Back in 2016, Donald Trump was elected president of the United States after one of the most unbelievable political campaigns in US history. Even on the day of the election, he was an underdog to win, with the opposition candidate Hilary Clinton leading the polls. The polls are rarely entirely accurate, but they usually give a strong indication of what is likely to happen. However, occasions such as the 2016 US election have shown that the prediction of the polls can be very different from the actual outcome of the event. How have the polls got it so wrong before, and could they get it wrong again?
Source: Unsplash
The Upcoming 2020 US Election
The polls for the upcoming 2020 election have Trump placed in a similar position to 2016. The Democrat nominee Joe Biden is leading the polls and Trump is once again the underdog. According another indicator of the likely outcome, the latest odds on the US presidential election, Trump is 15/8 to be re-elected with Biden being the bookies favourite at 2/5 as of 30th October. People may be more cautious of the polls and predictions this time and not be reading into them as much, but they do still offer some indication of how the general public is feeling towards the election. Experts are claiming the polls should be more accurate for the 2020 campaign, but where have they got it wrong in the past?
Representation in the Polls
A key area where the polls have previously got it wrong is how their predictions were weighted in terms of representation. In 2016, people without higher education were found to be heavily underrepresented in the polls. For example, 64% of white males that didn’t graduate from college are reported to have voted for Trump, however, the polls were unprepared for the sheer amount of people who fell into that category that would turn out to vote for Trump. He also gained the majority of votes from white women who had not completed college. The results led to claims that the polls had poorly represented the lesser-educated, as the majority of non-college graduates voted for Trump whereas most graduates voted for Clinton.
State Polling
The misrepresentation of how certain states were voting was another big problem for the pollsters back in 2016. Swing states such as Florida, Wisconsin and Michigan were tipped to go in the Democrats favour by the polls, however, they had severely underrepresented the Republican voters and all three states were won by Trump. One theory for how this happened is that many undecided voters ended up voting for Trump on the day of the election, as well as a lot of people privately voting for him without stating their intentions. It has since been widely accepted that changes needed to be made to the polling system to get a more accurate representation of specific state voter intentions.
Our electoral map, based on the latest pollshttps://t.co/1W7rRkeXh5 pic.twitter.com/ecinGEsKrx
— Political Polls (@PpollingNumbers) October 30, 2020
Pollsters have claimed to have improved their systems since the 2016 election in order to give a more accurate prediction. Despite this, the polls are still only predictions, so people are right to remain cautious about reading too much into them.